The present invention relates to an image detection system for a car, and in particular to a system for detecting and processing an image of a TV camera mounted on a car.
Heretofore, a system has been researched and developed in which a TV camera is mounted on a car as image detection means in order to find out obstacles on the road forward or backward of the car wherein an alarm is given to a driver, or a brake or a steering gear is automatically operated.
In this system, the TV camera can be mounted outside a car on one hand or inside a car on the other hand.
In the former case, it is disadvantageous in that raindrops in case of rain etc. or dust adhere to the surface of the lens of the TV camera, thereby disturbing a clear image.
A system for removing raindrops or dust which thus adheres to the surface of the lens of the TV camera is disclosed in Japanese Utility Model Application Laid-open No. 59-25528. However, this system is disadvantageous in that the lens of the TV camera should be rotated every time one desires to clean up raindrops or dust on the lens, which consumes time and disables an image taken meanwhile to be used.
Moreover, in the latter case where the TV camera is mounted outside a car, the TV camera per se is exposed to raindrops or dust so that the function thereof is easily deteriorated, resulting in a short life.
Therefore, the TV camera mounted inside a car is preferably to the one mounted outside a car.
However, it is also disadvantageous in that the image of a wiper for removing raindrops or dust intrudes into images detected by the TV camera. Therefore, if image processing is carried out as it is, a screen including the wiper image has no context with the adjacent screens so that the wiper image is to be detected as the nearest obstacle.
Thus, useless image processing with such an unavailable image, is carried out so as to cause the car erroneously to run in an automatic chase for an object.